Equimolar CO2 Absorption by Anion-Functionalized Ionic Liquids

Feb 1, 2010 - Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, Indiana 46556 .... Moreover, the results demo...
0 downloads 6 Views 268KB Size
Published on Web 02/01/2010

Equimolar CO2 Absorption by Anion-Functionalized Ionic Liquids Burcu E. Gurkan, Juan C. de la Fuente,† Elaine M. Mindrup, Lindsay E. Ficke, Brett F. Goodrich, Erica A. Price, William F. Schneider,* and Joan F. Brennecke* Department of Chemical and Biomolecular Engineering, UniVersity of Notre Dame, Notre Dame, Indiana 46556 Received November 2, 2009; E-mail: [email protected]; [email protected]

The discovery of materials that selectively and efficiently absorb CO2 from flue gases is essential to realizing practical carbon capture and sequestration (CCS). Ionic liquids (ILs) are promising in this regard because of their negligible vapor pressures, high thermal stability, and virtually limitless chemical tunability. Here we show how we can obtain extremely high capacity (up to one mole of CO2 per mole of IL) of CO2 in anion-functionalized ILs. CO2 has a large physical solubility in many ILs, and this solubility can be enhanced, for instance, by fluorinating the anion or cation components.1-3 Even with these improvements, however, Henry’s law constants are too large for practical CO2 separations from postcombustion flue gas. Taking a cue from the chemistry of aqueous organic amines reacting with CO2,4 amine functionalities can be added to ILs to introduce specific and tunable chemical reactivity with CO2. The first reported example contained an amine group on an imidazolium cation which reacted with CO2 in a manner similar to aqueous amines, by forming carbamate and ammonium ions in a stoichiometry of one CO2 to two amines,5 as shown in eqs 1 and 2.

the instability of the product dianion. These trends are most pronounced when the charge center and amine are in close proximity. Following our previous work,9 we performed electronic structure calculations at the B3LYP/6-311G++(d,p) level on isolated prolinate and methioninate anions reacting with CO2 and observed similar trends favoring 1:1 reaction. The net energies to form the prolinate and methioninate complexes are -71 and -55 kJ · mol-1, respectively. This difference reflects the diminishing effect of ring strain on the reactivity of the proline N.

Figure 1. Reaction schematics of CO2 with [P66614][Met] (top) and

[P66614][Pro] (bottom).

While CO2 uptake of ILs with amine-functionalized cations is much greater than is possible by physical absorption, this 1:2 stoichiometry is atom inefficient. The 1:2 reaction stoichiometry (eqs 1 and 2) effectively arises from amine deprotonation of a carbamic acid intermediate. A number of groups have explored other amine-functionalized ILs, including functionalized sultones,6 and amino acids anions with imidazolium7 and phosphonium8 cations. All systems have been assumed to produce 1:2 stoichiometry, although some of the evidence6 points to higher capacities and may warrant revisiting in light of the findings presented in this paper. Nonetheless, the question remains whether more favorable stoichiometry than one CO2 for two amines (1:2) can be achieved. We report here the synthesis and characterization of amino acidbased ILs, including trihexyl(tetradecyl)phosphonium prolinate ([P66614][Pro]) and methioninate ([P66614][Met]), that react with CO2 in a ratio of one CO2 per one amine (1:1 stoichiometry). Termination of the reaction sequence at (reaction 1) would result in a 1:1 stoichiometry and more atom-efficient use of the IL. We used ab initio calculations to explore the relationship between the position of the amine functional group and the relative energies of reactions 1 and 2.9 Calculations showed that tethering the amine to the cation favored the formation of the carbamate (reaction 2), reflecting the electrostatic stability of the zwitterions product, while tethering the amine to the anion favored the carbamic acid (reaction 1), reflecting † Present address: Departamento de Ingenierı´a Quı´mica y Ambiental, Universidad Te´cnica Federico Santa Marı´a, Valparı´so, Chile.

2116

9

J. AM. CHEM. SOC. 2010, 132, 2116–2117

[P66614][Met] and [P66614][Pro] were synthesized from reaction of [P66614][OH] with the corresponding amino acids. CO2 absorption was measured in a custom-built isochoric apparatus, containing an accurately calibrated stirred cell. From the initial and final pressures of CO2, the amount of CO2 absorbed was calculated with the ideal gas equation of state. The total equilibrium CO2 capacities of [P66614][Pro] and [P66614][Met] at 22 ( 1 °C are shown in Figure 2. Measurements shown were made in both a stoichiometric uptake apparatus and isochorically in an IR system. The data are in good agreement and the techniques are described in Supporting Information. Clearly, the CO2 capacity is significantly greater than one CO2 per two ILs and approaches the 1:1 stoichiometry expected from the theoretical arguments. There are two distinct portions of the isotherms: a large uptake at low pressures due to chemical complexation and further increase in capacity at higher pressures due to physical absorption. For both ILs, the vast majority of the uptake is due to chemical complexation, as discussed below. There is clearly a difference in the CO2 uptake curves for the [P66614][Met] and [P66614][Pro], with the [P66614][Pro] being much steeper at lower pressures, indicative of a higher heat of reaction. This is entirely consistent with the electronic structure results. Both isotherms evidently saturate at values slightly less than 1:1 stoichiometry. This deviation cannot be explained by residual halide impurities, and we see no evidence of partial carbamate formation (see below). The differences may arise from deactivation of amine groups via some mechanism at high CO2 loadings; further characterization is necessary to fully elucidate this behavior. As further evidence of the proposed reaction mechanism, we directly measured the heats of absorption and reaction of CO2 in [P66614][Pro] and [P66614][Met] using calorimetry at 25 °C and 2-3 bar. The experimental values are -80 kJ · mol-1 and -64 kJ · mol-1 10.1021/ja909305t  2010 American Chemical Society

COMMUNICATIONS

of CO2, respectively, in surprisingly good agreement with the reaction energies calculated for the gas-phase anions. The enthalpies were determined using a Setaram MicroDSCIII calorimeter, with CO2 introduced into a calibrated cell containing a known amount of IL. The heat flux needed to maintain the sample isothermally at 25 °C was measured, and knowing the CO2 capacity (Figure 2), this heat flux was converted to kJ per mol of CO2. The estimated experimental uncertainty is (5 kJ · mol-1 of CO2.

The solid lines in Figure 2 are fits of the data to an isotherm model derived for the reversible 1:1 reaction shown in eq 1, using the physical solubility Henry’s law constants. The excellent fits give additional support to the proposed chemistry.

Figure 3. IR spectrum of [P66614][Pro] before and after reaction with CO2

The inset shows disappearance of the N-H stretch.

Figure 2. CO2 absorption by [P66614][Pro] and [P66614][Met] at 22 °C. The lines are Langmuir model fits of the data included to guide the eye.

Furthermore, the 1:1 stoichiometry of the reaction of CO2 with [P66614][Pro] and [P66614][Met] has been confirmed by FTIR spectroscopy at 25 °C using a stirred Mettler Toledo model iC10 ReactIR system. The FTIR spectrum of [P66614][Pro] before and after reaction with CO2 (at 0.2 bar) is shown in Figure 3. There are three important features in the spectrum. First, the prolinate N-H stretch observed at 3290 cm-1 disappears when the CO2 reacts with the IL (Figure 3, inset). Concomitantly, the N-H bending around 1603 cm-1, which overlaps with the carbamate asymmetric stretch, decreases with reaction with CO2. Second, there is no evidence for ammonium formation (which would be present for a 1:2 stoichiometry). Specifically, one would expect ammonium bands between 3000 and 2800 cm-1 and combination bands in the 2800-2000 cm-1 region, if ammonium ion were present.10 This is clearly not the case. Third, there is a new band centered at 1689 cm-1, which corresponds to the new COOH moiety formed from the reaction of CO2 with the amine. Assignments are consistent with vibrational spectra obtained from the electronic structure calculations. The FTIR spectrum of [P66614][Met], shown in Supporting Information, shows similar features. The -NH2 group in the unreacted IL exhibits symmetric and asymmetric stretches centered at 3360 cm-1 with a shoulder at 3300 cm-1. This pair transforms into a single peak with the shoulder as CO2 reacts. Simultaneously, a new broad peak appears between 1760 and 1660 cm-1. Curve fitting applied in this region reveals two distinct CdO stretching environments: (i) 1688 cm-1 due to the COO- asymmetric stretch of the acid of the amino acid anion (symmetric stretch observed at 1469 cm-1) and (ii) 1718 cm-1 of newly formed COOH due to the proton transfer from the amine nitrogen. In addition, the FTIR spectra allow us to distinguish the fractions of physically dissolved and chemically reacted CO2, since the physically dissolved CO2 appears cleanly between 2370 and 2310 cm-1. A calibration plot was developed using tetraglyme (TG), 1-hexyl3-methylimidazolium bis(trifluoromethylsulfonyl)imide and 1-hexyl3-methylpyridinium bis(trifluoromethylsulfonyl)imide as standards with known CO2 gas solubilities.3,11,12 At 25 °C, we estimate physical solubility Henry’s law constants of 57 bar for [P66614][Pro] and 147 bar for [P66614][Met]. Thus, at 1 bar pressure of CO2 the physical uptake of CO2 at 25 °C by [P66614][Met] and [P66614][Pro] is less than 1% and 3% of the overall capacity, respectively.

The CO2 absorption capacities of a number of tetrabutylphosphonium amino acid ILs supported on high surface area silica gel have been reported,13 but in all cases any uptake in excess of 0.5 mol CO2 per mol of IL (i.e., 1:2 stoichiometry) has been attributed to physical absorption. Here, through a combination of theory, spectroscopy, calorimetry, and vapor-liquid absorption measurements, we have clearly shown that neat prolinate and methioninate ILs exhibit significantly higher absorption capacities, approaching 1 mol of CO2 per mol of IL. B3LYP calculations on glycinate and alaninate support a similar reaction mechanism and reaction energies intermediate between those of the anions studied here. In conclusion, we have shown that phosphonium-based amino acid ILs can react with CO2 in a 1:1 stoichiometry, achieving higher molar capacities than cation-functionalized ILs or even aqueous amine absorbents. Moreover, the results demonstrate that the location of functional groups (anion vs cation) in ILs is an additional degree of freedom available in the design of functionalized ILs for specific applications. Acknowledgment. We thank Edward J. Maginn for helpful discussions and Serena S. Mathews for the glass transition temperatures. This material is based upon work supported by the Department of Energy under Award Number DE-FC-07NT43091. Supporting Information Available: Materials, synthesis methods, and other experimental and computational details. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Anthony, J. L.; Anderson, J. L.; Maginn, E. J.; Brennecke, J. F. J. Phys. Chem. B 2005, 109, 6366–6374. (2) Baltus, R. E.; Culbertson, B. H.; Dai, S.; Luo, H. M.; DePaoli, D. W. J. Phys. Chem. B 2004, 108, 721–727. (3) Muldoon, M. J.; Aki, S. N. V. K.; Anderson, J. L.; Dixon, J. K.; Brennecke, J. F. J. Phys. Chem. B 2007, 111, 9001–9009. (4) Wolsky, A. M.; Daniels, E. J.; Jody, B. J. EnViron. Prog. 1994, 13, 214–219. (5) Bates, E. D.; Mayton, R. D.; Ntai, I.; Davis, J. H. J. Am. Chem. Soc. 2002, 124, 926–927. (6) Soutullo, M. D.; Odom, C. I.; Wicker, B. F.; Henderson, C. N.; Stenson, A. C.; Davis, J. H. Chem. Mater. 2007, 19, 3581–3583. (7) Fukumoto, K.; Yoshizawa, M.; Ohno, H. J. Am. Chem. Soc. 2005, 127, 2398–2399. (8) Fukumoto, K.; Kohno, Y.; Ohno, H. Chem. Lett. 2006, 35, 1252–1253. (9) Mindrup, E. M.; Schenider, W. F. In ACS Symposium Series; Seddon, K., Rogers, R., Plechkova, N., Eds.; American Chemical Society: Washington, D.C., 2009. (10) Silverstein, R. M.; Webster, F. X.; Kiemle, D. J. Spectrometric Identification of Organic Compounds, 7th ed.; John Wiley & Sons, Inc.: New York, 2005. (11) Anderson, J. L.; Dixon, J. K.; Maginn, E. J.; Brennecke, J. F. J. Phys. Chem. B 2006, 110, 15059–15062. (12) Sciamanna, S. F.; Lynn, S. Ind. Eng. Chem. Res. 1988, 27, 492–499. (13) Zhang, J. M.; Zhang, S. J.; Dong, K.; Zhang, Y. Q.; Shen, Y. Q.; Lv, X. M. Chem.sEur. J. 2006, 12, 4021–4026.

JA909305T J. AM. CHEM. SOC.

9

VOL. 132, NO. 7, 2010

2117